[WOPALX]

Originally Posted by shiv@vishnu

Yep. More than just a CAN specialist though, needless to say.

Shiv

He does a damn fine oil change too

[gian335]

Originally Posted by WOPALX

He does a damn fine oil change too

lol.....good info

[nj1266]

Originally Posted by Crazy AWD

Just plain wrong. The ECU has things like short and long fuel trim values that get modified based on variables like knock threshold.

Fuel trims have nothing to do with knock. Fuel trims are the ECU's method of adjusting AFR in closed loop. The front O2 sensor's main job in life is to keep the AFR at 14.7:1 to keep the CAT happy and keep emissions low. If the AFR drifts away from 14.7:1 in closed loop, the O2 sensor sends a signal to the ECU. The ECU in turn tell the injectors to add/subtract fuel to keep AFR at stoich.

For example, Let us say that your logged fuel trim was +8%. What does that mean? It means that the O2 sensor told the ECU that there was a lean condition detected in closed loop and the ECU added 8% more fuel to bring the the AFR back to 14.7:1 AFR.

The ECU does NOT THINK on its own. It reacts to feedback from the sensors by looking at 2d/3d tables that tell it what too do in thousands of contingencies. There is NO idependent thinking or adapting going on. The DME is simply reacting to waht the sensor data and the maps in the DME tell it to do. Change the sensors and the maps and the DME will give you different results.

You do mean the piezoelectric knock sensors (i hope). They are far from a "microphone."

Yes, that knock sensor is essentially a glorified microphone. Infact, knock detection is entirely primitive and reactive. There is no PROACTIVE knock detection at all. The car makes noise that the sensor picks up and sends to the ECU, the ECU filters the noise based on the 2d filter maps. If the conditions are met in the maps, then the ECU retards the timing to protect the engine from further knocking. It is REACTING to an already existing knock incident

[nj1266]

Originally Posted by adrian@vishnu

Interesting thread.

I think we need some new definitions of direction ignition control. Many people are referring to direct as the ability to run some number that can be set in a map somewhere regardless of whether this capability adds any value to the tuning process which it does not. Since the Procede can read DME igntiion timing via CAN, it could actually do direct ignition control going by this definition. All it would have to do is add one different calculation: CAS Offset = Procede Target - DME Actual timing. Then the Procede would have direct timing control. You could enter the timing you want it to run and the engine would run it. Would it actually achieve anything in terms of tuning

I do not define direct ignition control that way. Direct control is when the tuner has the ability to change the degree of timing retard/advance in the 3D maps that the DME uses to set ignition timing. That is direct control. What the interceptor is doing is "manipulating" the input from the sensors to achieve ignition timing that is different from what exists in the 3D maps.

When knock starts to occur again, they remove it again. There is likely to be multiple ignition base maps in the DME, and depending on the level of knock activity over a period of time, it may then decide that the fuel RON is lower and enter a base ignition map with lower timing which enables it to be a little closer to ideal timing from the base map. But it will then take a period of time with low knock activity to return to the more agressive base timing map. This may be the adaption people refer to on some tunes.

This describes well the high octane and low octane ignition and fuel maps. When knock is consistent and sever, the ECU switchs to the low octane maps to protect the engine from further damage. Low octane maps usually have richer AFRs and timing retard. When the knock conditions no longer exist, as per the signal from the knock sensor, then the ECU goes back to reading the high octane maps.

If this is the adaptation that BMW folks are refering to, then they are going to be in for a big surprise when their engine give up the ghost from detonation.

[munters]

Originally Posted by nj1266

If this is the adaptation that BMW folks are refering to, then they are going to be in for a big surprise when their engine give up the ghost from detonation.

and this is why control timing is essential. (direct or inderect who cares)
Oh by the way the interceptor is not the HW anymore. (I know you like the sneaky way)

[TheTwinz]

Originally Posted by nj1266

I do not define direct ignition control that way. Direct control is when the tuner has the ability to change the degree of timing retard/advance in the 3D maps that the DME uses to set ignition timing. That is direct control. What the interceptor is doing is "manipulating" the input from the sensors to achieve ignition timing that is different from what exists in the 3D maps.



This describes well the high octane and low octane ignition and fuel maps. When knock is consistent and sever, the ECU switchs to the low octane maps to protect the engine from further damage. Low octane maps usually have richer AFRs and timing retard. When the knock conditions no longer exist, as per the signal from the knock sensor, then the ECU goes back to reading the high octane maps.

If this is the adaptation that BMW folks are refering to, then they are going to be in for a big surprise when their engine give up the ghost from detonation.

interesting points you are bringing up, but you know you are arguing with someone that works at Haltech. The same company that made the XEDE for your car I'm guessing. So it's hard to tell who is right or wrong..

[The Cthulhu]

Originally Posted by nj1266

A DME is a computer; it is as smart as the input that you give it. The DME takes the information from the car's sensors (mainly, MAF/MAP/CPS), looks up 3D/2D tables from the rom (map) and sends back signals to the engine telling it to adjust spark ignition timing, fuel injectors firing and WGDC opening.

Some ECUs are more sophisticated than others, they have multiple tables that interpolate with each other based on different contingencies and some ECUs are simple. The DME happens to be a sophisticated ECU, but it DOES NOT THINK on its own. It DOES NOT ADAPT. It reacts to data from the sensors.

The interceptor (proceed) takes intercepts the data from the sensors and tells the ECU that it is getting a different signal than what it in the ECU map. The ECU says OK and sends a different signal to the sensors.

For example, say the ECU has in its timing map a command telling it to fire the spark plug @ 13* BTDC at 7000 rpm at WOT. The ECU gets a signal from the CPS sensor telling it to do just that, BUT the Proceed intercepts that signal and corrects it by -3* (CPS offset/correction). So now the ECU is told to fire the spark @ 10* BTDC @ 7000 rpm at WOT. The ECU thinks that it is following the map, but it is NOT, it is following what is coming from the sensors via the Proceed.

That is all there is to it. The ECU does not think, does not adapt and does nothing of what Mike is saying it does. The ECU takes the data in and spits different data out.

This, BTW, is NOT direct timing control. Direct timing control is when you flash the map in the ECU and the DME reads the new timing/fuel data. That, and only that is direct control.

For your second question, yes it is possible to determine detonation by logging data from the ECU. Each engine has a microphone attached to it. The microphone listens to the engine and sends the sounds to the ECU. The ECU has multiple filter maps that filter the sound and determines what is knock and what is not. If the ECU determines that a certain sound is knock, then it commands the CPS to retard iginition timing.

The ECU usually spits out the knock signal through the CAN/OBDII port and you can log it if you have the necessary logging equipment. Once you log it you will not that every time the knock signal goes higher, there is a corresponding timing retard.

When I ran an XEDE I was able to log that signal using an additional device on the harness. Everytime the signal increased the timing got pulled. Maybe Shiv can implement something like this on the Proceed.

+100

This info should be stickied, hah. Is there really such a thing as knock sensor noise or whatever people are claiming? Given the rate at which combustion cycles occur, I wonder if the knock sensor is spot on.

[Ilma]

Thanks for that explanation Adrian!

In the midst of all these claims and counterclaims, your's is one of the more informative and understandable posts that actually makes a lot of intuitive sense.

These timing control arguments remind me of when Shiv first started introducing throttle plate closure management concepts on these forums.

It was adamantly contested by some, that throttle plate management was best left to the DME to control in order to maintain BMW's built in safety systems. Of course, the fact that these safety systems were designed to work with stock boost levels was not taken into consideration in their arguments.

It only makes sense that as you increase power and boost, the stock design is no longer working with the same stock parameters. This is especially true at part throttle where the piggybacks cannot anticipate every possible throttle angle and load scenario and they fail to completely fool the DME with their offsets.

I think that in the end, the fact that throttle plate closure management was emulated by the competition speaks volumes to those who care to listen.

I run both tunes and like them both for different reasons.

In any case, yours is one of the few posts that is actually understandable to me and makes a lot of intuitive sense.

Last time I felt like this was when Shiv began explaining the why's and hows of excessive throttle closure by the DME and how this led to inconsistent performance......it made so much sense at the time.

[HighVoltage]

Originally Posted by The Cthulhu

+100
Is there really such a thing as knock sensor noise or whatever people are claiming? Given the rate at which combustion cycles occur, I wonder if the knock sensor is spot on.

With any given sensor there's always noise/signals that fall outside of the range you are interested in observing/monitoring. Thats the whole point of filters...

[lawdude]

Originally Posted by Ilma

Thanks for that explanation Adrian!

In the midst of all these claims and counterclaims, your's is one of the more informative and understandable posts that actually makes a lot of intuitive sense.

These timing control arguments remind me of when Shiv first started introducing throttle plate closure management concepts on these forums.

It was adamantly contested by some, that throttle plate management was best left to the DME to control in order to maintain BMW's built in safety systems. Of course, the fact that these safety systems were designed to work with stock boost levels was not taken into consideration in their arguments.

It only makes sense that as you increase power and boost, the stock design is no longer working with the same stock parameters. This is especially true at part throttle where the piggybacks cannot anticipate every possible throttle angle and load scenario and they fail to completely fool the DME with their offsets.

I think that in the end, the fact that throttle plate closure management was emulated by the competition speaks volumes to those who care to listen.

I run both tunes and like them both for different reasons.

In any case, yours is one of the few posts that is actually understandable to me and makes a lot of intuitive sense.

Last time I felt like this was when Shiv began explaining the why's and hows of excessive throttle closure by the DME and how this led to inconsistent performance......it made so much sense at the time.

Why is it that when Shiv posts an explanation/graphs or whatever, people come out nowhere to attack, but when Adrian posts an explanation you can just about hear crickets chirping.

[Sparky66]

Truth is hard to swallow.........especially when it's a JB3 pill.

[inthemiddle]

Originally Posted by munters

and this is why control timing is essential. (direct or inderect who cares)

+100 (+3295932587932)

[RambleJ]

This is very interesting information but to me it seems more like some..


I mean preemptive controlled timing or not, the JB3 is not a new tune. There are cars out there that have put 10's of thousands of miles on them and very little issues reported back..

[nj1266]

Originally Posted by TheTwinz

interesting points you are bringing up, but you know you are arguing with someone that works at Haltech. The same company that made the XEDE for your car I'm guessing. So it's hard to tell who is right or wrong..

I am not disagreeing with him on the substance of what he said. My disagreement is on the definition of "direct timing control." I define it as writing to the 3d timing maps in the ECU and letting the ECU read the degrees of timing retard/advance from the maps. Obviously, he defines it differently.

I am in agreement with him on everything else.

[OpenFlash]

Originally Posted by RambleJ

This is very interesting information but to me it seems more like some..


I mean preemptive controlled timing or not, the JB3 is not a new tune. There are cars out there that have put 10's of thousands of miles on them and very little issues reported back..

The vast majority of tuned cars on this forum probably have less than 20k tuned miles. And in a street car where less than 1% of that mileage is actually at WOT, that really isn't much of an accomplishment. It's only just now that we are seeing 15+psi tunes in the summer being the norm.

It's a bit said when reason, fact and logic get mistaken for scare tactics.

Shiv

[Mike@N54Tuning.com]

Hi Adrian,

Originally Posted by adrian@vishnu

* In the case of a system with CAS phase adjustment, the system can add this extra ignition retard. The shape of the curve is maintained (with respect to RPM), but the curve just has an offset applied to it. The actual ignition advance being run is the same as if the ECU had been correctly mapped for that boost level (hence why "direct" control is not required) assuming the tuner has mapped it well.

Assuming the above statement is correct, then the timing curve as seen by the DME with the CAS phase adjustment in place should reasonably match the factory stock boost curve. This is the same argument Shiv is making in the first post, by showing the stock timing log vs. tuned timing log. The problem is, his hand picked logs didn't match up. In fact they were 50% apart at times. They also did not cover the full power band. Having looked at several stock vs. V3 tuned logs now from 2000-7000rpm on 91 octane I can say they consistently do not match up. They are 6-7 degrees apart in spots during same day testing. So we are forced to accept that either Shiv has poorly tuned your system, or that CAS phase adjustments are failing to get the job done.

Originally Posted by adrian@vishnu

The knock sensor cannot measure when knock is almost happening. It can only measure when knock IS happening. There is however variable intensity of knock, and it can detect lower levels of knock than are likely to cause immediate damage, and can respond quick enough to prevent damage from prolonged knock. However the knock sensor cannot determine when conditions have changed to allow the ignition advance to be run without knock after it has retarded timing, so the algorithms slowly reintroduce the ignition that has been removed (slowly being seconds rather than milliseconds). When knock starts to occur again, they remove it again.

This is the same mechanism the DME uses to self compensate for octane variances in stock cars, and the same mechanism that we will show learns out the CPS offsets in the first place. These systems are always active whether or not the boost levels have been increased through a piggyback. As I've said the ECU has long term "octane" based adaptation, and short term "knock retard" based adaptation. The principle difference being in how aggressively it attempts to reintroduce advance.


Mike

[Mike@N54Tuning.com]

Originally Posted by Ilma

It was adamantly contested by some, that throttle plate management was best left to the DME to control in order to maintain BMW's built in safety systems. Of course, the fact that these safety systems were designed to work with stock boost levels was not taken into consideration in their arguments.

If you recall, that discussion was about the throttle plates role in boost control. It has since been shown with many logs, including some you've posted from both systems, that the throttle plate plays a critical role and must be left in the loop. The misunderstanding was that one side was rigging the MAP sensor feedback to prevent the throttle plate from closing when needed. Effectively taking control away from the DME. That assumption stemmed from claims of "zero throttle closure", which later turned out to be claims of "only moderate throttle closure". Subsequent logs showed the DME was in the loop, that throttle closure did occur when needed, and the argument was thus dropped.

Mike

[jonm42]

Unsubscribed.

[nj1266]

Originally Posted by Mike@N54Tuning.com

This is the same mechanism the DME uses to self compensate for octane variances in stock cars, and the same mechanism that we will show learns out the CPS offsets in the first place. These systems are always active whether or not the boost levels have been increased through a piggyback. As I've said the ECU has long term "octane" based adaptation, and short term "knock retard" based adaptation. The principle difference being in how aggressively it attempts to reintroduce advance.


Mike

What do you mean by "systems." There are no systems in the ECU. There are multiple 3d/2d maps. If you have access to these maps, then please show them to us. Please share with us the so-called "adaptation" maps. Post and image of these maps, please. Stop talking in the abstract and show us the adaptation maps in the ECU.

Every ECU map that I have looked at has a high octane and low octane timing and fuel maps. When the car is running w/o knock, the ECU uses data from the high octane maps. When the car starts to continuously knocks, the ECU interpolates to the low timing and fuel maps to preserve the engine. In these maps the AFR is richer and the timing is retarded.

Maybe the DME is different, but YOU will have to prove it by showing US the maps. You are the one questioning how the proceed works, not us.

So please get images of the maps from the DME and post them so we can all see this adaptation that you speak of.

Until then you have proved nothing and shown nothing.

[ZTUNER]

All these posts and basically here is what it boils down to.

Using the factory knock control to influence timing is bad practice regardless of how sophisticated the knock control system is. Fact is that its a "REACTIVE" system. In order for it to work, knock must occur. All Engine manufacturers use the factory knock control as a backup to their tune not as a primary tuning system.

Using CPS offset a hack that apparently works well enough. You are changing timing by changing the reference point. ie if the engine physically is at TDC, you tell the ecu its actually x deg atdc. This in effect reduces the effective ignition timing by x degrees. Fair enough. As for "learing it out" I don't think there is any "learning it out" since timing control is not open loop / closed loop like a/f. Some people say that the ecu/DME will keep adding timing till it senses knock and then back it off. I don't see that as being true. I think what they are confusing it with is the timing trim induced by knock. In other words, if you install a tune and there is no knock the ecu will not trim back any timing. If there is no trim to begin wit there will be no additional timing added.

All this being said, there is another facet to CPS offset. Offsetting CPS impacts ignition timing AND fuel injection timing. Since i did not know the effects of injection timing changes i contacted my engineer friends at ISUZU where i previusly worked. I was informed that injection timing affects on fuel consumption and if the change is large on the egts the engine sees. I was also informed that if the change is 4 - 5 deg the effect is not significant. So even though its a hack it works so long as huge changes are not made

In other words factory knock sensor based reactive timing control not good, cps based timing control, better but i would not call it "direct control on timing". In fact a piggy is a piggy is a piggy and will never be as good as a flash assuming the DME is cracked fully

Harry

[ZTUNER]

Originally Posted by Mike@N54Tuning.com

This is the same mechanism the DME uses to self compensate for octane variances in stock cars, and the same mechanism that we will show learns out the CPS offsets in the first place. These systems are always active whether or not the boost levels have been increased through a piggyback. As I've said the ECU has long term "octane" based adaptation, and short term "knock retard" based adaptation. The principle difference being in how aggressively it attempts to reintroduce advance.


Mike

It will be interesting to see how you show this. Looking forward to it. IMO for what its worth i hope you are correct 116 octane sonds interesting

Harry

[scalbert]

Originally Posted by Mike@N54Tuning.com

They also did not cover the full power band. Having looked at several stock vs. V3 tuned logs now from 2000-7000rpm on 91 octane I can say they consistently do not match up. They are 6-7 degrees apart in spots during same day testing. So we are forced to accept that either Shiv has poorly tuned your system, or that CAS phase adjustments are failing to get the job done.

I would appreciate data, and all of it. As I have said, I am open to review and discussions, just would like all data. Because, and as you know, I can get in the car and drive down the road and have significantly different timing curves depending on what gear I am in, have I been sitting in traffic, transitioning from on/off power, etc.

In other words, complete data including IAT, etc.

This is the same mechanism the DME uses to self compensate for octane variances in stock cars, and the same mechanism that we will show learns out the CPS offsets in the first place. These systems are always active whether or not the boost levels have been increased through a piggyback. As I've said the ECU has long term "octane" based adaptation, and short term "knock retard" based adaptation. The principle difference being in how aggressively it attempts to reintroduce advance.

Your suggestion would indicate that there has to be thousands of octane adaptation channel as it would be very RPM dependant. CPS offsetting is not consistent across the board nor does it occur at moderate load. Which the last statement would multiply the number of adaptation channels by a large amount of load variation. As such, the number of adaptation channels for timing control would dwarf the number of channels for fuel adaptation. That is quite a bit of data words to update and retain.

If not that complex, your assertion, if true, would be negated quickly in normal driving. Moderate loads induce little to no CPS offset so the DME would drift back to where it was before any offsetting occurred. But then the next time you got on it, the CPS is there to help.

Terry

Fixed.

And one last time as I see it has been ignored. Why did BMS state the JB3 had a proprietary timing control system that was a direct method which did not exist. This was only stated after CAS was found to not exist in the JB3. Was it was a lie, marketing, or what? Just put this one to bed so we can focus on the test data.